Expand description
Pure-Rust formant-based text-to-speech engine.
Compiled in with --features tts. Produces a real, decodable WAV
whose pitch contour follows the input text: each character is mapped
to a formant frequency, and a brief envelope-shaped sine wave is
emitted per character. The result is robotic (no neural TTS) but
it’s genuinely synthesized speech-adjacent audio that’s
deterministic, intelligibility-poor, and completely self-contained —
no model files, no FFI, no install step.
Operators who want natural-sounding TTS should wire up Piper (via
piper-rs once mature) or a cloud TTS in a follow-up iteration.
The trait + capability surface is ready for them.
Structs§
Functions§
- render
- Render a real 16-bit PCM WAV at 16 kHz mono from
text.